This paper deals with a new system of online Arabic handwriting recognition based on the association of beta-elliptic modeling extractor with a hybrid Time Delay Neural Network (TDNN) and Support Vector Machines (SVM) classifier. The beta-elliptic model proceeds by a segmentation of the handwriting trajectory into fragments called Beta strokes by inspecting the extremums points of the curvilinear velocity and extracting their corresponding static and dynamic profile proprieties. These features are used to train the Time Delay Neural Network which is able to represent the sequential aspect of the input data. The fuzzy outputs of this network are then used to train SVM in order to predict the correct label class. To evaluate our method, we have used a total of 25000 Arabic letters from the LMCA database. Experimental results demonstrate the effectiveness of our proposed method and show recognition rate reaching the 99.52%.
CITATION STYLE
Zouari, R., Boubaker, H., & Kherallah, M. (2017). Hybrid TDNN-SVM algorithm for online Arabic handwriting recognition. In Advances in Intelligent Systems and Computing (Vol. 552, pp. 113–123). Springer Verlag. https://doi.org/10.1007/978-3-319-52941-7_12
Mendeley helps you to discover research relevant for your work.